Transcript of IRC chat on C++ M2 release

[Andrew Borley <aj...@gmail.com>]

Hi all,

We just had a good IRC chat about the forthcoming Tuscany C++ M2 release.
Main things decided:
- We will try to get Release Candidate 1 out by end of this week/early next
- Items in RC1 have been decided & volunteered for, the list will be
available at http://wiki.apache.org/ws/Tuscany/TuscanyCpp/Tasks (when I've
done it)

Transcript below.

Cheers!
Andy

18:00:29<robbinspg>tick tock ... it's 18:00 BST
18:00:35<ajborley>Hi all - my clock says it's 18:00 - good to go?
18:00:43<ajborley>Pete beat me to it
18:01:02<robbinspg>I live slightly further east than you ;-)
18:01:03<jsdelfino>hey
18:01:27<robbinspg>Are you going to run this Andy?
18:01:41<ajborley>Yep, I'm happy to run this
18:01:49<ajborley>OK, so we've put a list of items up on the wiki
18:02:03<ajborley>http://wiki.apache.org/ws/Tuscany/TuscanyCpp/Tasks
18:02:34<ajborley>and there's a lot there, I expect quite a few things won't
make it if we want to get a first RC out by the end of this week
18:02:57<ajborley>So, the first question is - is an RC by the end of the
week a good idea?
18:03:15<ajborley>Any items that really should go in & will take more than a
week?
18:03:47<robbinspg>The first RC does not have to be perfect and I'd like to
see us have the infrastructure to make the release in place
18:04:14<robbinspg>so we know we can turn the handle and create a RC
18:04:16<jsdelfino>yes, I guess the first RC is there to allow us to polish
the build infra
18:04:22<ajborley>Yep sounds good - do you think the infrastructure can be
updated by the end of the week?
18:04:46<robbinspg>that means getting the licence stuff correct and a
working build and package on linux and windows
18:05:05<robbinspg>yes I think it's doable
18:05:08<ajborley>Yep - doesn't look too bad
18:05:37-->|slaws (i=chatzill@nat/ibm/x-4d3c867d81249199) has joined
#tuscany
18:05:52<jsdelfino>what do we need to do to the Linux build? is it good as
of today?
18:05:56<robbinspg>so I'm currently updating the licence info so mark that
one down to me
18:06:30<robbinspg>The linux build is ok but we need to make sure that make
dist works
18:06:40<ajborley>I think te linux build is mostly good - not sure what the
script that packages bin and src distros is good
18:06:59<robbinspg>and that the resulting distro contains all the right
stuff to enable ./configure, make in the samples
18:07:13<robbinspg>That took some work for M1
18:07:20<jsdelfino>ah ok, I can help test the linux build if it helps
18:07:45<ajborley>great - thanks
18:08:19<ajborley>Should we go through the list to decide if items should be
in or out?
18:08:24<robbinspg>Windows is built via a bat which basically runs the vc6
compile followed by copying/zipping
18:08:54<jsdelfino>are we building with vc6, 7 or 8 for this release?
18:09:04<ajborley>ok. How up-to-date are the vc7 files?
18:09:11<ajborley>sorry vc6
18:09:33<ajborley>I think we still have an issue with vc8
18:09:46<robbinspg>I vote we stay with vc6 for this releas
18:09:58<ajborley>I know vc7 is good, but I don't have a vc6 system
18:10:17<robbinspg>VC6 projects don't work at all at the moment but are
easily fixable
18:10:21<slaws>which version of vc7?
18:10:28<ajborley>Pete - could you do vc6 things for the release?
18:10:35<robbinspg>it;s the generating of the batch makefiles which makes
vc6 attractive
18:10:36<ajborley>I think mine is vc7.1
18:10:39-->|jmarino (n=jmarino@c-71-198-130-248.hsd1.ca.comcast.net) has
joined #tuscany
18:10:45<ajborley>Pete - agreed
18:10:54<jsdelfino>I suggest we stick to one, and maybe even take out the
other versions, it's a little confusing to have multiple versions of the
project files, some up to date, some outdated
18:10:56<ajborley>it's nice to have a command line build
18:10:59<slaws>if pete can generate the make can we use them?
18:11:19<ajborley>Yep, vc6 make worked with vc7 last time
18:11:27-->|gwinn (n=chatzill@slip32-106-64-7.lon.uk.prserv.net) has joined
#tuscany
18:11:33<slaws>k - sounds like a plan
18:11:38<robbinspg>I will fix up VC6 and generate the makefiles... and
oh  well I may as well do the distro packagingtoo
18:11:47<ajborley>:) cool
18:12:12<jsdelfino>ok, I'll try to get a VC6 install on my machine, I only
have vc8 now
18:12:36<jsdelfino>so we are using VC6 for this release right?
18:12:38<robbinspg>well you should be able to run the command line with your
VC7 I think
18:12:53<ajborley>jsdelfino - it would be good to try the command line build
with vc8
18:13:09<ajborley>+1 for vc6 build only
18:13:17<gwinn>I mainly use VC 7 so I can test that if no one else does.
18:13:25<jsdelfino>vc8 doesn't work at all for now, SDOs don't run, throwing
an incompatible list exception
18:13:28<robbinspg>ok. I'll post on dev when I have a working VC6 and
generated make
18:13:41<ajborley>great
18:13:56<jsdelfino>I'll try to get a VC6 install
18:14:04<robbinspg>So.... what samples will we include in the distro?
18:14:05<ajborley>So the other things we should work on this week are docs
and samples
18:14:14<ajborley>still quicker than me :)
18:14:17<robbinspg>jsdelfino ... I don;t think you need VC6
18:14:57<ajborley>samples- Calculator, PythonCalculator, RubyCalculator,
BigBank are all good at the moment
18:15:06<ajborley>not sure about SupplyChain
18:15:24<robbinspg>I think that's good. Let's leave SupplyChain out for now
18:15:50<ajborley>RubyBank? jsdelfino - do you know the status of that one?
18:15:58<robbinspg>How are we going to package the "optional" language
extensions - Ruby, Pyhton
18:15:58<jsdelfino>yes, it's working
18:16:18<jsdelfino>can we add a simple HelloWorld? what do you think?
18:16:35<robbinspg>Calulator is our HelloWorld
18:16:49<ajborley>I think Calculator is simple enough
18:17:00<jsdelfino>okay :) that's fine too
18:17:10<robbinspg>I don't think we should add anything else until we get
what is already there build/packaged/tested/documented
18:17:21<ajborley>yep, agreed
18:17:25|<--rfeng has left freenode (Read error: 110 (Connection timed out))
18:17:46<ajborley>So samples are Calculator, PythonCalculator,
RubyCalculator, BigBank, RubyBank
18:17:54<robbinspg>+1
18:18:02<jsdelfino>any PHP?
18:18:02<ajborley>We need to get these documented
18:18:23<ajborley>There's a question about the PHP extension - should that
be included?
18:18:47<ajborley>It only does services without refereneces or properties at
the mo
18:18:48<robbinspg>does it work?
18:19:13<slaws>it does what it does - just not very extensive
18:19:20<jsdelfino>if we include the PHP extension, we need a sample for it
18:19:36<slaws>i made the calculart sample work with it
18:19:55<slaws>but that possibly got factored out in the reorg
18:20:08<slaws>i think i made a patch
18:20:17<slaws>but that might not be relevant now
18:20:21<robbinspg>so are we going to include the following language
extensions: cpp, Python, Ruby, PHP
18:21:03<ajborley>I'm happy with them, but we will need a PHPCalculator
sample
18:21:04<slaws>if you go with it we would have to mark it "PHP patial"
18:21:07<robbinspg>We need to document how to build/install each of htese
18:21:10<ajborley>yep
18:21:15<jsdelfino>I think that if we want to include the PHP extension it
should support references and simple properties
18:21:25<robbinspg>Shall we leave PHP out of RC1 then
18:21:40<slaws>y - i think thats wise
18:21:41<robbinspg>we can add it in later after proving the build etc works
18:21:52<ajborley>ok, I'm good with that
18:22:08<jsdelfino>yes, we can leave it out of RC1, and keep the option to
add it to RC2 or 3 :)
18:22:14<ajborley>Maybe get it in after RC1
18:22:39<ajborley>So what docs do we need?
18:22:57<ajborley>How to build SCA/SDO
18:23:06<ajborley>How to build/run the samples
18:23:15<robbinspg>I was determined to get no further than RC3 for M1,
that's why we got RC3, RC3a, RC3b ;-)
18:23:31<jsdelfino>how to build the various extensions (the pre-req software
etc.)
18:23:40<robbinspg>all good so far
18:23:40<ajborley>How to create a component for C++/Python/etc
18:24:20<ajborley>Design docs?
18:24:39<ajborley>e.g. "how to write an extension"?
18:24:39<robbinspg>The minimum is how to build/install core and the
extensions plus how to build and run the samples
18:24:46<ajborley>yep - for RC1
18:24:48<robbinspg>no design docs
18:24:53<ajborley>agreed
18:24:59<robbinspg>no time for RC1
18:25:17<jsdelfino>maybe we can use the doc of the samples to describe the
PM for Python and Ruby
18:25:33<jsdelfino>yes, no design docs for RC1
18:25:52<robbinspg>yes to samples describing the PM
18:26:15<ajborley>+1 also
18:26:56<robbinspg>what about stdcxx? If we get optionally building SDO with
this do we document it? SCA will not have this option for this release
18:26:57<ajborley>OK - I'll go through the wiki & put [RC1] next to the
things we've decided to go into it
18:27:27<ajborley>what's teh status with that? Can that be ready y end of
the week?
18:27:43<gwinn>If we don't document stdcxx for SDO then there's not much
point having it.
18:27:45<ajborley>y -> by
18:28:38<ajborley>gwinn - could that be documented by the end of the week?
18:28:52<jsdelfino>don't we have SDO working with stdcxx? what is there to
document?
18:29:03<robbinspg>That was my question ;-) We should have the optional
stdcxx build for SDO but not for SCA
18:29:09<jsdelfino>yes
18:29:16<ajborley>yep
18:29:30<robbinspg>and if you build SDO with stdcxx it won't work with SCA
18:29:44<robbinspg>and that's fine for now
18:29:51<jsdelfino>yes I think that's ok for now
18:30:05<ajborley>agreed
18:30:07<jsdelfino>we have to go step by step
18:30:17<ajborley>so the release packages we're doing are:
18:30:20<gwinn>We need a small amount of documentation to explain how to
build with stdcxx
18:30:37<ajborley>sdo on stdcxx for Linux bin
18:30:41<robbinspg>yes... thanks for volunteering gwinn ;-)
18:30:49<gwinn>ok
18:31:02<ajborley>sdo for Linux bin
18:31:42<ajborley>I give up - SDO with/without stdcxx on Linux/Windows
18:31:42<robbinspg>not sure we should distro the SDO with stdcxx as bin
18:32:15<ajborley>yep. if it doesn't work with sca will having it there just
muddy the water
18:32:20<ajborley>?
18:32:33<jsdelfino>can we have the stdcxx option documented, so people can
build it from source?
18:32:34<robbinspg>we have SDO src lin and win + SDO bin lin and win (vc6)
18:32:53<ajborley>yep. with teh docs for using stdcxx
18:32:58<robbinspg>agreed
18:33:02<jsdelfino>+1
18:33:15<ajborley>and then SCA src lin & win + SCA bin lin & win
18:33:17<gwinn>I back jsdelfino. stdcxx + SDO only from source
18:33:28<robbinspg>now for SCA we have src package for lin and win
18:33:48<robbinspg>and... what do we put in the bin???
18:33:53<robbinspg>all the extensions
18:34:00<robbinspg>or jsut core cpp , ws
18:34:02<ajborley>everything I would say
18:34:27<jsdelfino>yes, unless it breaks on machines that do not have the
prereq software, but I don
18:34:30<jsdelfino>'t think it does
18:34:32<ajborley>the runtime won't need what it doesn't use
18:34:39<robbinspg>well we could package the language extensions separately
18:34:59<ajborley>I think a single package would be better
18:35:11<jsdelfino>I'm ok with both :)
18:35:12<robbinspg>agreed... let's try for that and see what breaks
18:35:20<ajborley>yep :)
18:36:06<ajborley>what about MacOSX? there was a patch for that, but I don't
know what the status is
18:36:17<ajborley>I also don't have access to any mac machines
18:36:19<robbinspg>I'll do Mac OSX
18:36:32<robbinspg>so long as someone buys me a new PowerBook
18:36:47<jsdelfino>for MacOSX the first question is: is Axis2C shipping bins
on MacOSX?
18:36:50<ajborley>:) do we want RC1 on OSX? Could anyone volunteer?
18:37:12<ajborley>Good point - I doubt it
18:37:22<robbinspg>let's get RC1 out and see if the resulting src release
can be built on Mac
18:37:36<jsdelfino>we should ask oisin what he wants to do, since he
contributed that patch and tested it
18:37:44<robbinspg>our one linux bin distro will be????
18:38:10<ajborley>the RHEL3 one seemed to work in lots of places last time
18:38:11<robbinspg>I built on RHEL3 last time
18:38:23<ajborley>it worked on my fedora 5 box
18:38:32<jsdelfino>worked on my rhel4
18:38:43<robbinspg>ok so let's stick with that
18:38:59<ajborley>dunno if anyone's got a non-redhat linux distro around?
18:39:01|<--rfen1 has left freenode (Read error: 145 (Connection timed out))
18:39:15<jsdelfino>do u have the right prereqs on your rhel3 for ruby,
python and php?
18:39:25<robbinspg>not yet!
18:39:38<jsdelfino>:)
18:39:39<robbinspg>I couldn't sort out PHP
18:39:44<robbinspg>but yes to the others
18:39:59<robbinspg>but that is fine for RC1
18:40:03<ajborley>yep
18:40:03<jsdelfino>ok good
18:40:24<jsdelfino>going through the extensions, I think we won't have a
javascript :)
18:40:49<ajborley>Dependencies - Axis 0.93? 0.94 is out soon but i don't
think this week
18:40:53<robbinspg>So here is what I think I am doing: Licence headers, VC6
build, windows distro
18:41:16<robbinspg>0.93 for now
18:41:25<jsdelfino>I tested the axis2c trunk / close to 0.94 yesterday and
it worked, I just had to change one char* ot a const char* in our code
18:41:34<jsdelfino>but 0.93 is good for now, yes
18:41:38<ajborley>cool
18:42:14<ajborley>the libxml2 stuff for SDO is the same as last release
right?
18:42:17<jsdelfino>REST will work but only REST POST, GET doesn't work well
with 0.93
18:42:33<robbinspg>jsdelfino to look at Linux build and distro?
18:42:38<gwinn>libxml2 - yes.
18:43:13<robbinspg>ajborley (Release Manager) to take it easy?
18:43:21<jsdelfino>yes I can help with the Linux build and distro, but I'm
using RHEL4
18:43:24<ajborley>:) yep
18:43:36<ajborley>I'll look at docs & samples
18:43:49<robbinspg>RHEL4 should be fine for getting the automake to generate
the correct stuff
18:44:04<gwinn>Am I supposed to volunteer now?
18:44:08<robbinspg>yes
18:44:32<gwinn>OK. stdcxx docs & linux implementation. What else?
18:44:48<robbinspg>I'm happy doing the SDO build/distro as it's nice and
easy
18:44:58<ajborley>I guess I have to do the package signing stuff?
18:45:08<gwinn>OK. Just show me how for next time.
18:45:15<ajborley>and build the actual packages?
18:45:34<slaws>chaps, i'm holding back just at the moment as i'm knee deep
in PHP stuff
18:45:40<ajborley>(that go up on people.apache.org?)
18:45:45<robbinspg>yes... when we have the mechanism in place we need to run
the complete build/package then sign, generate md5
18:45:48<jsdelfino>jsdelfino will be at ApacheCon next week so it'll be
convenient if we need help from the Apache incubator, Axis2C or stdcxx teams
18:46:20<robbinspg>great
18:46:42<robbinspg>so Andy will now update Wiki to reflect what we just
agreed to??
18:47:01<ajborley>yep - I'll put RC1 & names next to the items
18:47:05<jsdelfino>one more question:
18:47:17<jsdelfino>are we going to support complex types in Python?
18:47:18<jsdelfino>:)
18:47:26<ajborley>not for RC1 :)
18:47:50<jsdelfino>we could just map them to strings if it's simpler, but I
think we need something for RC2 or 3...
18:48:14<jsdelfino>an extension with no XML complex type support is very
limited
18:48:23<ajborley>true
18:48:24<robbinspg>Let's see what the sample/extension loks like in an RC1
first
18:48:34<jsdelfino>ok
18:49:03<ajborley>ok - any more items?
18:49:07<jsdelfino>yes
18:49:23<jsdelfino>we need to doc the level of spec we support and
limitations
18:49:29<jsdelfino>a section in our README I guess
18:49:35<jsdelfino>I can help with that
18:49:43<ajborley>+1 for README
18:49:47<robbinspg>yes, could you invent 3 more days in the week
18:49:54<jsdelfino>do we need similar doc for SDO?
18:49:56<ajborley>what is the situation? are we at 0.95?
18:50:15<robbinspg>0.95 for C++ C&I
18:50:31<robbinspg>latest Assembly draft is 0.96
18:50:37<robbinspg>do we match that?
18:51:24<robbinspg>I guess SDO support is stated as 2.01 spec same as last
time as there is no new draft available
18:51:40<gwinn>SDO deviations from spec would be quite short unless you
include bugs.
18:52:02<gwinn>eg we don't do type safe interface (and that's in teh spec)
18:52:16<ajborley>I think there are bits of 0.96 assembly that we don't
support
18:52:19<robbinspg>yes... that's already doc'd
18:52:23<jsdelfino>we match 0.96 with some limitations
18:52:34<ajborley>e.g. full property support
18:52:48<robbinspg>ok... if you know what they are then we should doc them
18:52:49<jsdelfino>y
18:53:13<ajborley>ok, in the README
18:53:27<ajborley>cool. anything else?
18:53:31<robbinspg>Jiras
18:54:02<robbinspg>Are there any high priority issues that need fixing
18:54:15<ajborley>I don't think there are in SCA
18:54:28<gwinn>Nothing in SDO that can be done in the time.
18:54:34<robbinspg>also... we should raise Jiras for ALL check-ins from here
on in
18:55:03<jsdelfino>+1
18:55:09<robbinspg>makes it easier to see where we are
18:55:20<ajborley>yep - shall we each do that or shall I raise a load based
on this conversation?
18:55:40<jsdelfino>both maybe :)
18:55:40<robbinspg>thanks for volunteering ;-)
18:55:46<ajborley>ok - will do
18:56:14<ajborley>I'll assign them too :)
18:56:52<jsdelfino>I am in the middle of a change to simplify the target
deploy structure (remove the req. for configuration/ and packages/), I'll
create a JIRA for that
18:57:06<ajborley>Will that get into RC1?
18:58:12<jsdelfino>I'd like to, I'm almost done with this, checking in later
today
18:58:15<robbinspg>yes... I hope so
18:58:18<ajborley>great
18:58:48<robbinspg>ok. I need to go and open some wine. Anythin else to
discuss?
18:59:00<ajborley>Not from me
18:59:08<gwinn>ditto
18:59:23<jsdelfino>a little early for wine here, 11am :)
18:59:43<jsdelfino>thanks guys, have a good evening...
18:59:57<ajborley>never too early! right - see you on the dev list
18:59:57<robbinspg>bye.
19:00:03<ajborley>bye
19:00:07<ajborley>& thanks all
19:00:30<jsdelfino>bye
19:00:39<gwinn>bye

Re: Transcript of IRC chat on C++ M2 release

[Oisin Hurley <oh...@iona.com>]

>> Quick clarification Andy - will you be shifting axis version
>> before Milestone? Or saving it for after?
>
>
> It depends when 0.94 comes out I guess -  if it's soon then  
> shifting is
> probably a good idea (especially with some of the fixes that are  
> planned for
> 0.94), but I don't think we should delay M2 specifically for 0.94.  
> Does that
> sound OK?

Sounds fair enough to me, thanks!

  cheers
   --oh

---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org

Re: Transcript of IRC chat on C++ M2 release

[Andrew Borley <aj...@gmail.com>]

On 10/4/06, Oisin Hurley <oh...@iona.com> wrote:

>
> Quick clarification Andy - will you be shifting axis version
> before Milestone? Or saving it for after?


It depends when 0.94 comes out I guess -  if it's soon then shifting is
probably a good idea (especially with some of the fixes that are planned for
0.94), but I don't think we should delay M2 specifically for 0.94. Does that
sound OK?

Cheers

Andy

Re: Transcript of IRC chat on C++ M2 release

[Oisin Hurley <oh...@iona.com>]

> That would be good - we decided near the end of the IRC chat to have a
> dependency on Axis2C 0.93 as we weren't sure exactly when 0.94  
> would be out
> and we'd like our first RC by the end of this week. Saying that,  
> Sebastien
> put in a tiny code change which made Tuscany work with the latest  
> Axis2C
> trunk, so the 0.94 release will hopefully work with Tuscany M2.

Ok, well I notice their code freeze was lifted there today, so I'll
go see what I have to do to get a build together on 0.94. The 0.93
build did have some unit test failures on MacOS, it will be interesting
to see if any of the current spate of updates has fixed these.

> I have no idea - I'm not sure if the Tuscany site is the best place  
> to host
> an Axis2C MacOSX binary, but we could certainly point to one on  
> another
> (perhaps independent?) site. Another option is to point to/provide the
> patches to enable OSXers to build Axis2C and then Tuscany.

If I get a build and test clean, I will do some volunteering to
do the builds and tests for the Axis2C guys in the interim so
maybe they will host the bin build.

> Cheers Oisin, that's great - I guess we won't have the time to get  
> it into
> RC1, but we may be able to sort it out for a later RC.

Quick clarification Andy - will you be shifting axis version
before Milestone? Or saving it for after?

  cheers
   --oh

---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org

MacOS X build, was: Transcript of IRC chat on C++ M2 release

[Jean-Sebastien Delfino <js...@apache.org>]

Andrew Borley wrote:
> Hi Oisin,
>
> On 10/4/06, Oisin Hurley <oh...@iona.com> wrote:
>>
>> Hi guys,
>> Sorry I couldn't make the IRC - had to run off to a night class.
>>
>> > 18:36:06<ajborley>what about MacOSX? there was a patch for that,
>> > but I don't
>> > know what the status is
>> > 18:36:17<ajborley>I also don't have access to any mac machines
>> > 18:36:19<robbinspg>I'll do Mac OSX
>> > 18:36:32<robbinspg>so long as someone buys me a new PowerBook
>> > 18:36:47<jsdelfino>for MacOSX the first question is: is Axis2C
>> > shipping bins
>> > on MacOSX?
>>
>> I can get a MacOSX build done for 0.94, but as you can see at [0], the
>> priority on the patch is low. Looking at the current set of Jirae that
>> are going in there it seems they have more pressing concerns.
>>
>> This brings up a question from me wot is ignorant of Apache devoirs
>> in term of infrastructure support. Since I'm mostly doing Eclipse
>> work, I'm used to having heaps of donated build machines that will
>> do nightlys etc. I guess that this is not the case at Apache and
>> that binary builds are done by project committers - or the most
>> excellent release person :)
>
>
> Yep, that's right. I assume at the moment that there is no-one working on
> Mac OSX in Axis2C - I'm sure they would be very happy to have Mac OSX
> coverage if someone can provide the build & testing work.
>
> I will update the Axis patch to work with the current pre-0.94 base
>> and see will they take it (it's not a new feature imho).
>
>
> That would be good - we decided near the end of the IRC chat to have a
> dependency on Axis2C 0.93 as we weren't sure exactly when 0.94 would 
> be out
> and we'd like our first RC by the end of this week. Saying that, 
> Sebastien
> put in a tiny code change which made Tuscany work with the latest Axis2C
> trunk, so the 0.94 release will hopefully work with Tuscany M2.
>
> If they don't, would it be a horrible gaffe to post an 'independent'
>> set of binaries and headers as a supplemental download on the
>> Tuscany site? I hope asking that question wasn't of itself a
>> horrible gaffe :)
>
>
> I have no idea - I'm not sure if the Tuscany site is the best place to 
> host
> an Axis2C MacOSX binary, but we could certainly point to one on another
> (perhaps independent?) site. Another option is to point to/provide the
> patches to enable OSXers to build Axis2C and then Tuscany.
>

+1, if there's no better option then I think we should at least provide 
the patches and instructions to build Tuscany and its dependencies on 
MacOSX from source. Many people in open source communities now use 
MacOSX and without a MacOSX build they won't be able to work with 
Tuscany C++ at all. With all the effort that Oisin already put into it, 
I think it makes sense to at least publish these instructions and make 
clear that people can download our source distro and follow these 
instructions to build on MacOS X.

>> 18:36:50<ajborley>:) do we want RC1 on OSX? Could anyone volunteer?
>> > 18:37:12<ajborley>Good point - I doubt it
>> > 18:37:22<robbinspg>let's get RC1 out and see if the resulting src
>> > release
>> > can be built on Mac
>> > 18:37:36<jsdelfino>we should ask oisin what he wants to do, since he
>> > contributed that patch and tested it
>>
>> Consider me volunteered for build and test - I would consider that
>> there is more important work to be done for RC1, so I'm not going
>> to be insistent on it's inclusion.
>
>
> Cheers Oisin, that's great - I guess we won't have the time to get it 
> into
> RC1, but we may be able to sort it out for a later RC.
>

Great Oisin, thanks! As I said above I think it's really important to 
have a MacOS X build for all the MacOS X open source developers out there.

> Thanks!
> Andy
>


-- 
Jean-Sebastien


---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org

Re: Transcript of IRC chat on C++ M2 release

[Andrew Borley <aj...@gmail.com>]

Hi Oisin,

On 10/4/06, Oisin Hurley <oh...@iona.com> wrote:
>
> Hi guys,
> Sorry I couldn't make the IRC - had to run off to a night class.
>
> > 18:36:06<ajborley>what about MacOSX? there was a patch for that,
> > but I don't
> > know what the status is
> > 18:36:17<ajborley>I also don't have access to any mac machines
> > 18:36:19<robbinspg>I'll do Mac OSX
> > 18:36:32<robbinspg>so long as someone buys me a new PowerBook
> > 18:36:47<jsdelfino>for MacOSX the first question is: is Axis2C
> > shipping bins
> > on MacOSX?
>
> I can get a MacOSX build done for 0.94, but as you can see at [0], the
> priority on the patch is low. Looking at the current set of Jirae that
> are going in there it seems they have more pressing concerns.
>
> This brings up a question from me wot is ignorant of Apache devoirs
> in term of infrastructure support. Since I'm mostly doing Eclipse
> work, I'm used to having heaps of donated build machines that will
> do nightlys etc. I guess that this is not the case at Apache and
> that binary builds are done by project committers - or the most
> excellent release person :)


Yep, that's right. I assume at the moment that there is no-one working on
Mac OSX in Axis2C - I'm sure they would be very happy to have Mac OSX
coverage if someone can provide the build & testing work.

I will update the Axis patch to work with the current pre-0.94 base
> and see will they take it (it's not a new feature imho).


That would be good - we decided near the end of the IRC chat to have a
dependency on Axis2C 0.93 as we weren't sure exactly when 0.94 would be out
and we'd like our first RC by the end of this week. Saying that, Sebastien
put in a tiny code change which made Tuscany work with the latest Axis2C
trunk, so the 0.94 release will hopefully work with Tuscany M2.

If they don't, would it be a horrible gaffe to post an 'independent'
> set of binaries and headers as a supplemental download on the
> Tuscany site? I hope asking that question wasn't of itself a
> horrible gaffe :)


I have no idea - I'm not sure if the Tuscany site is the best place to host
an Axis2C MacOSX binary, but we could certainly point to one on another
(perhaps independent?) site. Another option is to point to/provide the
patches to enable OSXers to build Axis2C and then Tuscany.


> 18:36:50<ajborley>:) do we want RC1 on OSX? Could anyone volunteer?
> > 18:37:12<ajborley>Good point - I doubt it
> > 18:37:22<robbinspg>let's get RC1 out and see if the resulting src
> > release
> > can be built on Mac
> > 18:37:36<jsdelfino>we should ask oisin what he wants to do, since he
> > contributed that patch and tested it
>
> Consider me volunteered for build and test - I would consider that
> there is more important work to be done for RC1, so I'm not going
> to be insistent on it's inclusion.


Cheers Oisin, that's great - I guess we won't have the time to get it into
RC1, but we may be able to sort it out for a later RC.

Thanks!
Andy

Re: Transcript of IRC chat on C++ M2 release

[Oisin Hurley <oh...@iona.com>]

Hi guys,
Sorry I couldn't make the IRC - had to run off to a night class.

> 18:36:06<ajborley>what about MacOSX? there was a patch for that,  
> but I don't
> know what the status is
> 18:36:17<ajborley>I also don't have access to any mac machines
> 18:36:19<robbinspg>I'll do Mac OSX
> 18:36:32<robbinspg>so long as someone buys me a new PowerBook
> 18:36:47<jsdelfino>for MacOSX the first question is: is Axis2C  
> shipping bins
> on MacOSX?

I can get a MacOSX build done for 0.94, but as you can see at [0], the
priority on the patch is low. Looking at the current set of Jirae that
are going in there it seems they have more pressing concerns.

This brings up a question from me wot is ignorant of Apache devoirs
in term of infrastructure support. Since I'm mostly doing Eclipse
work, I'm used to having heaps of donated build machines that will
do nightlys etc. I guess that this is not the case at Apache and
that binary builds are done by project committers - or the most
excellent release person :)

I will update the Axis patch to work with the current pre-0.94 base
and see will they take it (it's not a new feature imho).

If they don't, would it be a horrible gaffe to post an 'independent'
set of binaries and headers as a supplemental download on the
Tuscany site? I hope asking that question wasn't of itself a
horrible gaffe :)

> 18:36:50<ajborley>:) do we want RC1 on OSX? Could anyone volunteer?
> 18:37:12<ajborley>Good point - I doubt it
> 18:37:22<robbinspg>let's get RC1 out and see if the resulting src  
> release
> can be built on Mac
> 18:37:36<jsdelfino>we should ask oisin what he wants to do, since he
> contributed that patch and tested it

Consider me volunteered for build and test - I would consider that
there is more important work to be done for RC1, so I'm not going
to be insistent on it's inclusion.

  cheers
   --oh

[0] https://issues.apache.org/jira/browse/AXIS2C-287

---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org